THE EFFECT OF ENERGY DEPLETION

ON ECONOMIC GROWTH

John D. Sterman

WP 1130-80 May 1980

System Dynamics GroupAlfred P. Sloan School of Management

Massachusetts Institute of Technology

D-3216

The financial support of the

Office of Analytical Services,

Policy and Evaluation,

U.S. Department of Energy

is gratefully acknowledged.

2

D-3216

TABLE OF CONTENTS

I. The Importance of Energy-Economy Interactions 4

II. The Need for an Integrating Framework 7

II. 1 Common Assumptions of Energy Supply Models 7

II.2 Characteristics of Energy-Economy Models 11

III. Structure of the Model 18

IV. Simulation Results 25

IV. 1. Reference Run 27

IV.2 The Importance of Substitution 43

IV.3 The Importance of Delays 48

V. An Example of Energy Policy Analysis 54

VI. Conclusion 59

APPENDIX 61

REFERENCES 75

3

D-3216

THE EFFECT OF ENERGY DEPLETION ON ECONOMIC GROWTH

I. THE IMPORTANCE OF ENERGY-ECONOMY INTERACTIONS

The 1970s might well be called the decade of energy. In 1970,

domestic production of oil peaked; natural gas production peaked in 1972.

Both oil and gas output stood lower at the end of the decade than at its

start. During the '70s, OPEC rose to dominance in world energy markets,

coal was again hailed as king but failed to ascend the throne, and nuclear

power generated a critical mass of safety, envirornmental, and ethical

opposition which has brought its development to a virtual standstill.

Depletion of energy resources has emerged as one of the major

problems facing the world for the remainder of the century. At the same

time, the economy of the United States has not fared well. The '70s saw

economic growth falter from the 3.7% per year rate of the '50s and '60s to

2.7% per year. The nation experienced the deepest recession since the

Great Depression, rising unemployment, and the most severe peacetime

inflation in U.S. history. While not all the nation's economic woes can be

traced to energy, the impact of energy on the economic health of the nation

is undeniable. The unemployment, factory shutdowns, and hardship caused by

the OPEC embargo of 1973, natural gas shortages of 1976, coal strike of

1978, and gasoline shortages of 1979 all testify to the direct importance

of energy in maintaining a modern industrial economy. But energy also

affects the economy in more subtle ways: energy prices have outpaced

inflation for most of the decade, adding to inflationary pressures; growing

4

D-3216 5

capital requirements for energy production threaten investment in other

sectors of the economy; financial markets strain to accomodate

petrodollars; and the dollar falls in value as OPEC prices rise, leading to

still higher prices.

The health of the economy depends critically on energy. Economic

policy can no longer be made without considering its impact on the energy

crisis. Energy policy can no longer be made without regard to its economic

repercussions. Yet there is little agreement on the nature, strength, and

relative importance of the myriad interconnections between energy and the

economy. A framework is needed to integrate in a consistent and realistic

manner the long-term effects of energy depletion and rising energy costs on

investment, economic growth, unemployment, the standard of living, and

inflation.

This paper seeks to demonstrate the feasibility of such an

integrating framework. The framework consists of a system dynamics model

of the national economy (the ENECON model). The purpose, structure, and

assumptions of the model will be described. The relevance of the approach

for policy analysis will be demonstrated by evaluating the impact of energy

depletion on economic activity, investment, and the standard of living. In

particular, the following questions will be addressed:

D-3216 6

o What is the relationship between GNP and energy use?

o Can depletion of energy resources cause or worsen an-economicdownturn?

o How do lags in the substitution process affect economicperformance?.

o Is shortage of capital (financial and physical) likely toconstrain the expansion of energy supplies?

III

D-3216 7

II. THE NEED FOR AN INTEGRATING FRAMEWORK

Since 1973 many energy models have been constructed and numerous

studies have been undertaken to address the energy crisis. In what way

does the ENECON model differ from other energy models? Why is there a need

for an integrating framework to examine interactions between energy and the

economy?

II.1 Common Assumptions of Energy Supply Models:

Because the energy crisis was initially perceived as a supply

problem, early energy models focused primarily on supply and treated the

energy sector in isolation from the rest of the economy.* Table 1 presents

several common assumptions of energy supply models. While these

assumptions simplify analysis of future energy supplies and prices, they

are quite unrealistic:

GNP Growth is Exogenous:

Imagine the effect on GNP of a full embargo of all imported oil

(50% of consumption). Massive unemployment, shutdown of factories,

widespread bankruptcy, and political turmoil would follow, whether the

price system or rationing were used to clear the energy market. It is the

fear of just such an embargo that led to the creation of the Strategic

Petroleum Reserve. Yet in energy supply models GNP would be unaffected.

* An early exception is found in White (1971).

D-3216

TABLE 1

COMMON ASSUMPTIONS OF ENERGY SUPPLY MODELS

GNP GROWTH IS EXOGENOUS

CAPITAL COSTS OF NEW TECHNOLOGIES ARE EXOGENOUS

INVESTMENT IN

OTHER SECTORS

ENERGY UNCONSTRAINED BY INVESTMENT NEEDS OF

OF THE ECONOMY

INTEREST RATES ARE EXOGENOUS

INFLATION IS UNAFFECTED BY ENERGY PRICES AND AVAILABILITY

WORLD OIL PRICES ARE UNAFFECTED BY DOMESTIC IMPORT

REQUIREMENTS OR ENERGY POLICIES

Representative models include: PIES, FOSSIL1 and FOSSIL2,DFI, Livermore Energy Policy Model, Baughman-Joskow, andBechtel Energy Supply Planning Model.

8

III

D-3216 9

Capital Costs of New Technologies are Exogenous:

The capital costs of technologies such as coal liquefaction and

gasification, shale oil, and nuclear power have risen over the past decade

in real terms, often (as in the case of shale oil) faster than the price of

oil itself! Such behavior is not surprising since these technologies are

extremely energy- and capital-intensive. When energy costs rise, the costs

of producing steel and concrete rise as do the costs of fuel and

feedstocks. Yet these interdependencies are ignored by the supply models.

Investment in Energy Unconstrained by Investment Needs of Other

Sectors of the Economy:

As the economy grows, the "pie" of national output to be divided

among consumption, investment and government spending grows. But because

the pie, though growing, is limited, the increasing capital requirements of

the energy sector can only be satisfied at the expense of investment in

some other sector or at the expense of consumption. The assumption that

GNP is unaffected by the energy sector implies investment in non-energy

sectors and consumption can be maintained in the face of rising energy

sector needs, and therefore that the energy sector cannot increase its

share of the pie. No capital constraints exist in the supply models:

energy industries do not compete for investment against the rest of the

economy. Thus the supply models let the economy have its "pie" and eat it

too.

D-3216

Interest Rates are Exogenous:

Growing investment requirements of the energy sector put upward

pressure on interest rates. As interest rates rise, investments earning

only marginal returns will become unprofitable and will be crowded out,

reducing capital investment in other sectors. Energy projects on the edge

of profitability may become unprofitable, requiring government subsidies

for completion.

Inflation is Unaffected by Energy Prices and Availability:

In the supply models inflation is exogenous. Yet energy prices

rose faster than the consumer price index during the '70s, adding to

inflationary pressure. As the price of energy rises, the cost of producing

every good and service in the economy rises (including the costs of

energy production). Higher costs are passed into prices, possibly

triggering a wage-price spiral, reducing the standard of living, and adding

to the demand for credit and to government deficits; each of these adds to

inflationary pressure. Energy projects with long lead times are

particularly vulnerable to inflation. Regulated industries such as

electric utilities are also vulnerable because allowed rates of return are

based on historical costs, not replacement costs.

World Oil Prices are Unaffected by Domestic Import Requirements or

Energy Policies:

Large oil imports worsen the balance of payments and weaken the

dollar; as the dollar drops in value OPEC raises the price of oil to

compensate for the loss in purchasing power, creating a vicious circle of

10

III

D-3216 11

escalating price hikes and devaluation. If import requirements were

reduced, the vicious circle would be weakened and OPEC's grip on the world

oil market would loosen. Indeed, one major benefit of reduced imports may

be a softening in OPEC prices (DOE, 1979). But the coupling between OPEC

and the domestic energy system is only one-way in the supply models.

Summary:

The assumptions commonly made in energy supply models are not only

contradicted by recent trends but ignore potentially important

interdependencies among energy costs, capital formation, growth, and

inflation. The results of these models are therefore called into question,

and their utility for energy policy-making compromised.

II.2 Characteristics of Energy-Economy Models:

In the past few years, a variety of models have been constructed

to address the deficiencies noted in the supply-oriented models.* These

"energy-economy" models are primarily designed to assess the impact of

energy availability and price on economic growth. The models employ a wide

range of techniques, from linear programming to econometrics to

input/output analysis. However, the models fail to capture important

energy-economy interactions in a manner that recognizes the imperfect

nature of energy markets and the complex dynamics underlying macroeconomic

change. Table 2 lists some of the properties of the major energy-economy

models. These properties are discussed below.

* For a comprehensive summary of research into Energy-Economy interactionsto 1979, see DOE/EIA (1979).

D-3216

Table 2

CHARACTERISTICS OF ENERGY-ECONOMY MODELS

EQUILIBRILU ORIENTATION [1] [3] [4] [7]

NO REBALANCING OF FACTORS OR FINAL DEMAND [1] [3] [5] [7]

IMPACT OF THE ECONOMY ON THE ENERGY SECTOR IGNORED [1] [5]

PHYSICAL AND FINANCIAL FLOWS NOT CONSERVED

PRODUCTION INDEPENDENT OF ENERGY

[5] [6]

[5]

[3] [5] [6] [7]

WORKFORCE PARTICIPATION IS EXOGENOUS

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[1] [2] [4] [5] [6] [7]

Elephant-Rabbit, Hogan and Manne (1977)

ETA-MACRO, Manne (1977)

PILOT, Parikh (1976)

Hudson-Jorgenson LITM, Hudson and Jorgenson (1978)

DRI, Hull (1979) and Forrester and Mass (1979)

Wharton Annual Energy Model, WEFA (1978)

Reister-Edmonds Two Sector Model, Reister and Edmonds (1977)

LACK OF ROBUSTNESS

12

III

D-3216 13

Equilibrium Orientation:

Energy markets have been heavily regulated for decades,

introducing market imperfections by distorting and biasing the information

that normally would cause consumers and firms to respond to the growing

scarcity of energy. Thus, the assumption in many models of a market-

determined equilibrium is not defensible.

Yet even without government-imposed imperfections, the energy

system does not conform to the usual assumptions of equilibrium economics.

The energy system is characterized by extremely long delays. It takes

upwards of 15 years to build a nuclear power plant, 5 to 8 years to build

coal and shale plants, 3 to 5 years to develop offshore oil and gas.

Automobiles last 5 to 10 years. Houses, factories, and offices last 20 to

50 years or more. Modern settlement patterns, transportation networks,

agricultural and industrial techniques, and lifestyles evolved over even

longer periods and will respond only slowly to changes in energy prices and

availability. On the government side, the record of the past few years

demonstrates the existence of substantial delays in developing and

implementing energy policies.

Thus, it will take decades for the economy to fully reconcile

itself to the end of the era of cheap energy. In the interim period, while

adjustments to the new order are incomplete, the impact of high energy

prices may be more severe than in the long run. Models (including

Elephant-Rabbit, PILOT, Hudson-Jorgenson, and Reister-Edmonds) that only

yield an equilibrium snapshot of the economy at some future date are quite

D-3216

limited in their ability to realistically assess the consequences of

delays, capital turnover, and long lead times. The question "where are we

going?" is crucial, but equally so are the questions "can we get there from

here?" and "how long will it take?"

No Rebalancing of Factors or Final Demand

As the price of energy rises relative to prices of other factors

of production and other goods, use of the less expensive inputs will

increase and energy use will be reduced. Firms will substitute labor and

energy-efficient capital for energy-intensive capital. Consumers will

favor products with less energy content over more expensive energy-rich

items. These rebalancing effects may have significant impacts on

unemployment, wages, relative prices, the mix of goods and services

produced, and lifestyles. Models that ignore these effects exclude a

potentially powerful set of forces which may mitigate the effect of higher

energy prices.

Impact of the Economy on the Energy Sector Ignored:

Several of the models (such as the Elephant-Rabbit model) treat

capital and investment in energy production exogenously or have exogenous

energy prices. Yet capital production (primary metals, construction,

shipbuilding, etc.) is an energy-intensive sector of the economy and thus

highly dependent on the price and availability of energy. Further, the

ability of the economy to generate enough capital to satisfy the growing

needs of the energy sector is a key question. By treating capital

exogenously, these models assume away an important economic dimension of

the energy problem.

14

Ill

D-3216 15

Physical and Financial Flows not Conserved:

It is extremely important to conserve both physical factors

(capital, labor, energy) and financial flows (wage payments, profits,

taxes) so the full effects of depletion and energy policies are captured.

For example, a government subsidy for production of synthetic fuels must

either raise tax rates, reduce government transfers or expenditures, or

increase the deficit; each of these will have a detrimental impact on the

economy, possibly offsetting the beneficial impact of the subsidy. The

DRI and Wharton models ignore the budgetary implications of energy

policies.

Production Independent of Energy:

In several models, especially econometric models such a the DRI

model, energy is not required for the production of goods. In such a model

a reduction in energy input has no direct effect on output. But as

mentioned above, a full embargo of imported oil would result in widespread

disruptions and a large reduction in output. Models that ignore the

physical dependence of production on energy cannot be used to evaluate the

impact of energy on economic performance.

Lack of Robustness

Many energy-economy models are not robust in the sense that they

behave implausibly under extreme conditions. For examnple, in a recent

application of the DRI model (Hull 1979), the long-run elasticity of oil

demand (based on historical data), was 0.2. Thus, to reduce oil

consumption by 50% (enough to eliminate imports), oil must rise to $800 per

D-3216

barrel. Yet at that price, the annual oil bill would be $2.3 trillion, a

sum larger than GNP in 1979.* A model must behave plausibly outside the

range of historical experience if it is to be useful for policy analysis

since changing conditions and policies may result in previously unobserved

behavior. A model that only makes sense in historical ranges of behavior

can only be used to analyze historical policies.

Workforce Participation is Exogenous:

Labor is a major determinant of GNP. To treat labor input exogen-

ously implies GNP is largely determined without regard to the effects of

energy prices and availability on wages, incomes, unemployment, and life-

style changes which may affect the decision to participate in the work-

force. The energy crisis could foster new low-consumption lifestyles in

which people spend more time working for themselves and less in the

workforce. Models with exogenous labor force participation include the

Elephant-Rabbit, ETA-MACRO, Hudson-Jorgensen, DRI, and Reister-Edmonds.

* Assuming current oil consumption of approximately 16 MMBD at an average

price of $25 per barrel. GNP in 1979 was approximately $2.1 trillion.

16

III

D-3216 17

Summary:

The models currently available are inadequate representations of

the complex interdependencies between energy and the economy. To assess

the importance of these interactions properly, a model should capture the

feedbacks between the energy sector and the rest of the economy. It should

generate and allocate the productive resources of the economy endogenously

and represent the shifting balance between these resources. Important

delays and disequilibrium effects should be included. Physical and

financial flows must be conserved. It should not rely on exogenous time

series to generate its behavior. Finally, the model should be robust under

extreme conditions. Such a model can then begin to shed light on the

magnitude of the problem, the severity of potential tradeoffs, and the

efficacy of policy initiatives.

D-3216

III. STRUCTURE OF THE MODEL

A system dynamics model (the ENECON model) has been constructed

for the analysis. Based on the System Dynamics National Model, the model

is designed to capture the decision-making structure of the economy.* The

model represents a first framework for the analysis of energy-economy

interactions and will undergo further development. The simulations

presented here are intended to illustrate the use of the model and to point

out important areas for future model development. An overview of the model

structure appears in Figure 1. There are five basic sectors: production,

household, financial, government, and OPEC.

The production sector actually consists of three individual

production sectors: one each to produce energy, capital plant and

equipment, and consumer goods and services. A sector represents many

firms, each producing similar types of output. Each production sector has

the same basic or generic structure which constitutes a theory of the firm.

Investment and capacity acquisition, hiring and firing policy, pricing,

production scheduling, and financial management are all represented

explicitly. Though the three producing sectors share a common structure,

each sector is calibrated to represent a particular type of output (such as

consumer goods or energy). For example, the initial ratio of capital to

labor, energy intensity of production, and debt-equity ratio of the sectors

differ to capture their differing production technologies and financial

characteristics.

* For a description of the system dynamics methodology, see Forrester(1961); for a description of the National Model see Forrester, Mass, andRyan (1976).

18

III

D-3216 19

Figure I OVERVIEW OF MODEL STRUCTURE

D-3216

Each production sector employs three factors of production--labor,

capital, and energy. These inputs are combined through a production

function to yield the output of the sector. Similarly, the household

sector uses four factors to generate "utility" or well-being: labor,

capital (representing housing), consumer goods and services, and energy

(which is necessary to utilize household capital and goods). The three

production sectors and the household interact through orders for output,

shipments, and payments, all of which are represented separately. Figure 2

shows the physical flows of energy, capital, and consumer goods that

connect the production and household sectors.

The energy sector differs from the other sectors in two important

respects. First, energy production depends on non-renewable resources. As

the easily exploited resources are consumed, the productivity of the

energy sector declines. Second, if the energy sector is unable to meet the

total demand for energy, it will attempt to import the difference from the

OPEC sector.

The OPEC sector uses its revenues to purchase current output of

the domestic economy and to invest in the productive assets of the economy,

that is, to purchase claims to future output.

At this stage, the energy sector produces a single type of energy.

There is no distinction between different fuel types. Interfuel substitu-

tion, therefore, is not treated.

20

III

D-3216 2121

Figure 2

THE PRODUCTION AND HOUSEHOLD SECTORS. . ._ , . ._ _ _ _ . ._ . .

D-3216

The financial sector is the source of external financing for

production sectors and the household, and is the repository for savings.

The financial sector allocates available funds for investment among

competing demands for credit, sets interest rates, and balances the

aggregate portfolio of the economy as the relative profitability of the

production sectors changes. For example, if return on investment is higher

than average in a sector, the sector will be able to attract funds for

investment at the expense of other sectors. The financial sector also

includes the Federal Reserve and can influence the money supply through

open market operations.

The government sector sets tax rates, makes transfer payments, and

determines fiscal and monetary policy. Energy policies are also the

province of the government sector.

Table 3 summarizes the major variables in the model, dividing them

into those endogenous to the model, those exogenous, and those that are

outside the scope of the study. The last category is of particular

interest. Energy depletion and the subsequent adjustment of the economy is

a process requiring decades. By comparison, adjustments in inventories of

consumer goods and energy are short-term phenomena and can thus be ignored

without compromising the conclusions of the study. As a result, the model

does not deal with short-term business cycle behavior.* International

trade, except energy imports from OPEC, is excluded for simplicity. While

* Mass (1975) presents a theory of business cycles based on interactions.between inventory management and employment policies.

22

III

D-3216 23

Table 3

SUMMARY OF MAJOR VARIABLES

ENDOGENOUS

GNP

CONSUMPTION

INVESTMENT

SAVINGS

STANDARD OF LIVING

PRICES (NOMINAL AND REAL)

CONSUMER PRICE INDEX

WAGES (NOMINAL AND REAL)

LABOR FORCE PARTICIPATION

EMPLOYMENT

UNEMPLOYMENT

INTEREST RATES

MONEY SUPPLY

DEBT (PRIVATE AND GOVERNMENT)

TAX RATES

ENERGY PRODUCTION

ENERGY DEMAND

ENERGY IMPORTS

ENERGY RESOURCES

EXOGENOUS

POPULATION

OPEC PRODUCTION

REAL OPEC PRICE

D(CLUDE

INVENTORIES

ENVIRONMENTALCONSTRAI NTS,COSTS

NON-ENERGYRESOURCES'

INTERNATIONALTRADE (EXCEPTWITH OPEC)

DISTRIBUTIONALEQUITY

TECHNOLOGICALPROGRESS

D-3216

the exchange rate, international capital flows, and terms of trade affect

the economy, they are not of primary importance in assessing the impact of

energy on the economy. The nature of environmental constraints (such as

clean air standards) and possible scarcity of non-energy resources (such as

water for synfuel development) are quite important but, again, for

simplicity have been excluded from the study. Finally, the effect of

rising energy prices on distributional equity is an issue with important

political consequences; though it deserves study, it is beyond the scope of

the research presented here.

The appendix describes the nature and determinants of equilibrium

in the model and reviews initial conditions.

24

III

D-3216 25

IV. SIMULATION RESULTS

The ultimate goal of developing the framework described above is

policy analysis: the model will be used to evaluate the efficacy and

potential side effects of various energy policies. A necessary prelude to

policy analysis, however, is the identification and exploration of the

channels through which energy influences the economic health of the nation.

The strength of the various relationships must be estimated, and potential

tradeoffs evaluated.

Several simplifying assumptions have been introduced to aid the

investigation. First, the model is initialized to represent an economy in

a stationary equilibrium. Population growth and technological progress

have been excluded. Second, since the focus is on investment, output, and

the standard of living--all real quantities--the money supply is kept

constant, preventing long-run inflation. Third, the tax rate (personal and

corporate) is held constant, and the government is assumed to balance its

budget. Fourth, OPEC is assumed always to satisfy the demand for energy

imports (embargoes and supply interruptions are excluded). OPEC is assumed

to recirculate its revenues by purchasing consumer goods from the domestic

economy. Finally, the real costs of energy production are assumed to

increase smoothly but at an increasing rate as depletion occurs until the

price of a backstop technology is reached.* In the model, the backstop

technology is available in unlimited quantities and without any extra

* Depletion paths and the concept of a backstop technology are discussedin Nordhaus (1973) and Solow (1974).

D-3216

development or construction delays. Such an assunption will underestimate

the effects of depletion on the economy, especially in the short run.

A caveat must be entered here. The real world does involve a

growing economy, inflation, government deficits, supply interruptions, and

development delays. But the purpose is not to reproduce the actual history

of the economy or predict its future, but to understand how energy

depletion affects the performance of the economy. By setting up a model of

the economy incorporating these assumptions, controlled experiments can be

performed which reveal the impact of depletion alone. As understanding

improves, simplifying assumptions can be relaxed.

In the simulations no attempt has been made to reproduce the

year-by-year performance of the economy. While the initial conditions,

parameters, and configuration of the model correspond to those of the

United States during the post-war period, the nunerical values shown in the

simulation should not be interpreted as predictions or forecasts. Rather,

the initial equilibrium should be used as a point of comparison for

evaluating the simulation results. Subsequent tests should be contrasted

against the reference run discussed below. It is the relative magnitude of

the effects, the overall behavior mode of the economy that is important.

The uncertainty and confusion over the effect of depletion on economic

performance is so great that great precision is unwarranted.

26

D-3216

IV. 1 Reference Run

In the first simulation the resource base is assumed equivalent to

about 50 years of consumption at the original usage rate. As depletion

proceeds, the productivity of the energy sector declines, reaching a final

level only one-quarter as great as the original. Thus, the backstop tech-

nology is assumed to be four times more expensive in real terms than the

original cost of the non-renewable resource. Figure 3 shows the returns to

oil drilling activity in the U.S. The return, in barrels per foot drilled,

declined by more than a factor of four up to the late 1960s, indicating the

factor of 4 increase in costs due to depletion assumed in the simulations

is a conservative estimate.

2

-0N,

0"OZ!,00110

20Cr

0 2 4 6 8 10 Iz 14 10 io ,u

Cumulative Exploratory footage (108 ft)

Figure 3. Returns to oil drilling activity in the U.S.1860-1967. Source: Hubbert (1969) p. 181.

27

h

D-3216

The simulation results are presented in Figure 4. Depletion

becomes significant in the 20th year; conversion to the backstop occurs

shortly before the 80th year. Real gross national product declines to

nearly three-quarters of its initial value and then recovers somewhat,

reaching equilibrium after 120 years at only 90% of the initial level. The

consumer price index (CPI) rises steeply in the period of depletion to 130%

of its initial level, then declines from the peak to equilibrium at 120%.

I I I I I ' I. ,f r I I

RE· f'e

I I I I I I

PRICE,Iuraw I I I I

U.,

U.,'N9

a CARS s a

Figure 4. The Reference Run: Real GNP and the Consumer Price Indexas a percent of their initial values; unemploymentas a percent of the workforce.

28

Ill

D-3216 29

Unemployment declines initially but then rises to a peak of more than 15%;

in equilibrium unemployment has returned to the initial 5% rate.

Two aspects of the behavior are notable. First, real GNP is

reduced in equilibrium, indicating a reduction in the material standard of

living. Second, the adjustment to equilibrium required more than half a

century once the backstop was reached. The impact of depletion on GNP,

unemployment, and the price level is much more severe during the adjustment

period than in the final equilibrium.

The essence of depletion is an increase in the effort required to

find, develop, and produce each additional barrel of oil or cubic foot of

gas. The real costs of energy production rise, or equivalently, the

productivity of the energy sector declines. Since the economy has only

limited productive resources at its command, (ultimately a reflection of a

limited supply of labor), a decline in energy sector productivity

necessarily implies a decline in the productivity of the economy as a

whole. The nation becomes poorer.

Figure 5 illustrates the basic relationships between energy

resources, production, investment, and consumption that produce the

behavior in Figure 4. Non-energy production, or GNP, depends on inputs of

energy as well as capital and labor. Given the initial capital stock,

labor force, and energy consumption in the energy sector, energy production

D-3216

ENERGY RESOURCES

rkIrn )nratEa f

tNtKi(Y rKI-JUA I UN +

N

CONSIHPTION *I NON-ENERGY . I NVESTMENTPRODUCTION

O r

+

PRODUCTIVE RESOURCESALLOCATED TO ENERGY

i+

INVESTMENT IN ENERGY

j/+

+INVESTMENT INNON-ENERGY

ff

PRODUCTIVERESOURCESALLOCATED TONON-ENERGY

t

Figure 5. Relationships between energy resources, production,investment, and consumption.N.B. The causal loop diagrams presented here aresimplified representations of the actual model structure.

30

D-3216

will fall. As depletion reduces the productivity of the energy sector,

energy production will tend to fall. If investment patterns do not change

and productive resources are not reallocated to energy production, GNP and

consumption will fall as energy production declines. Energy imports do not

provide a way out of the dilemma. If imports are used to make up the

difference between production and demand, the standard of living must still

decline as current output or claims to future output are transferred to

OPEC in payment for oil.

If, on the other hand, capital and labor are diverted into the

energy sector to offset the effect of depletion, GNP and consumption will

decline as these resources are drained from the other sectors of the

economy. In terms of Figure 5, investment in energy will increase at the

expense of investment in the non-energy sectors essential for growth.

Thus, depletion inevitably reduces the material standard of living

below what it would otherwise have been, regardless of how the resources of

the economy are allocated. The question becomes not "does energy affect

economic growth?", but "how strong is the effect?"

Other analysts, especially (Hogan 1978), (Hogan and Manne 1977),

and (Manne 1978) have argued the severity of the impact depends strongly on

the relative importance of energy as an input to production. Since energy

historically accounted for only 5% of the costs of production throughout

the economy, an increase in the energy sector's factor requirements should

have a small effect on the economy. Hogan and Manne liken the energy

31

D-3216

sector to a rabbit in an "elephant and rabbit stew": just as adding

another rabbit to the stew will still leave elephant as the dominant

flavor, so an increase in the energy sector's share of the productive

resources of the economy should have a small effect on overall economic

performance. Hogan and Manne go on to argue that given the relative size

of the energy sector, the magnitude of the impact of depletion depends

strongly on the ability of firms and consumers to substitute labor and

energy-efficient capital for energy. If other inputs can easily be

substituted for energy then the impact should be negligible. At the other

extreme, if no substitution is possible, that is, if energy and output move

in lockstep, the impact will be severe. In terms of Figure 5, the relative

size of the energy sector and the potential for substitution determine the

necessary degree of reallocation of resources between energy and non-energy

production. The severity of the tradeoff increases as the ability to

substitute declines.

In the reference run (Figure 4), the long-run elasticity of

substitution between energy and other inputs was assumed to be unity. If

energy accounts for roughly 5% of the costs of production throughout the

economy, (a value consistent with pre-1973 experience), then an elasticity

of one implies a 1% reduction in energy use can be offset with only a .06%

increase in capital and labor, leaving output unchanged. Economnetric

studies have generated widely divergent estimates of the elasticity. Hogan

and Fromholzer (1977) and Hogan and Weyant (1978) report several attempts

to estimate the elasticity; Berndt and Wood (1977) review principal

estimation efforts through 1977. Berndt and Wood (1979) sumnarize the

32

D-3216 33

controversy by noting that the estimates supporting substitutability (an

elasticity near unity) generally derive from cross-sectional data or

engineering process analysis, while estimates supporting complementarity

(an elasticity near zero) generally derive from aggregate time-series data.

The reference run employs an elasticity of unity, representing the high end

of the range; a later simulation (Section IV.2) eliminates substitution of

capital and labor for energy to test the sensitivity of the results to

uncertainty in the value of the elasticity.

*

oco o

LA '0LA

Cm =U11

EnMN

I I

0. . .. . .. . . . . .O

I I I I

YEARS a o 0liu yV

Figure 6. Reference Run: Real Energy Price (ratio to initial value)and Energy/GNP ratio (percent of initial value).

III

.

I

I

I

II

I

I

I

. . . .

D-3216 34

Figure 6 shows the real energy price and energy/GNP ratio from the

base run. The real price of energy rises as depletion occurs, reaching a

maximum 6 times its initial level before dropping back to an equilibrium

almost five times greater. Note that the real price of energy rose by a

factor of five while the productivity of the energy sector declined by only

a factor of four. The extra increase in energy price is due to indirect

factor requirements of the energy sector: as real energy prices rise, real

capital costs also increase, raising the costs of energy production still

further (see below).

As the price of energy rises, substitution efforts reduce the

energy intensity of the economy, and the ratio of energy use to GNP begins

to drop. In equilibrium, the ratio stands at one-fifth of its initial

value, reflecting the unit elasticity of substitution.

According to the Elephant-Rabbit model (Hogan and Manne 1977), an

elasticity of unity implies higher energy costs have virtually no impact on

GNP. Yet the simulation results show a 10% decline in real activity

resulting from a four-fold decline in energy sector productivity. The

explanation for the difference in conclusions lies in the differences

between the structure of the two models. In the Elephant-Rabbit model,

capital and labor are exogenous. Labor, capital, and energy are not

required to produce capital, and energy is not needed to produce energy.

Further, because capital is exogenous, investors are assumed to supply

capital services for production regardless of the rate of return earned on

that capital.

D-3216 35

In the ENECON model, each sector uses labor, capital, and energy

as factors of production. The capital sector needs capital; the energy

sector needs energy. Plant and equipment are required to produce equipment

and erect buildings; energy is used to mine coal, drill for oil, and

especially to convert primary fuels such as oil to electricity. Investors

are assumed to require a target rate of return on their investment and will

raise prices to cover that target return and their operating costs.

In the reference run, the elasticity of substitution in each

sector is one. If capital and labor were exogenous, as in the Elephant-

Rabbit model, the magnitude of the GNP reduction would also be small. But

capital is not exogenous. The desired capital stock in each sector depends

on the cost of capital plant and equipment relative to the costs of labor

and energy. Because energy is required to produce capital, the cost of

plant and equipment rises relative to labor. Unlike the Elephant-Rabbit

model, capital prices rise to cover the higher energy costs and to insure

the target rate of return for investors. As a result, the goods, capital,

energy, and household sectors become relatively less capital intensive and

more labor intensive. Since the labor supply is fixed, the increase in

labor intensity results from a reduction in output beyond the reduction

induced directly by higher energy costs.

In addition, three mutually reinforcing relationships worsen the

impact still further (Figure 7). First, the energy sector requires energy

to produce. With energy five times more expensive, the energy sector cuts

back on its energy use and attempts to substitute even more capital and

III

ENERGY REQUIREMENTSOF ENERGY SECTOR +

+

ENERGY REQUIREMENOF CAPITAL SECT(

+

ITS)R

+CAPITAL REQUI-REMENTSOF ENERGY SECTOR

FOR CAPITALT

wIT\

CAPITAL REQUIREMENTSOF CAPITAL SECTOR

Indirect effects of increasing energy sectorcapital requirements.

D-3216 36

Figure 7.

N.

III

D-3216 37

labor. Thus, the labor intensity of the energy sector is increased

directly while the increase in capital intensity indirectly raises the

labor requirements of the energy sector by requiring the capital sector to

expand and use more labor. Again, since the labor supply is fixed,

equilibrium output.is reduced even further below the Elephant-Rabbit level.

Second, the capital sector uses capital as a factor of production,

creating another reinforcing relationship. Higher energy costs raise

capital production costs and make capital even more expensive relative to

labor. Thus, the capital sector becomes still more labor intensive,

further reducing equilibrium output.

Finally, the energy sector requires capital to produce and the

capital sector requires energy. As depletion increases the capital

intensity of the energy sector, the indirect energy requirements of the

energy sector also grow. Because energy becomes so much more expensive,

capital becomes less attractive, further increasing the labor intensity of

the energy and capital sectors and reducing equilibrium output. Table 4

shows the changes in output, real prices relative to the real wage, and

factor intensities in each sector caused by depletion in the reference run.

Capital intensity in the economy as a whole increases less than labor

intensity. Capital intensity remains unchanged in the capital sector and

actually declines in the goods sector. In the energy sector, capital

intensity increases only 88% as much as labor intensity.

D-3216

Real Output

Real Price ofOutput Relativeto the Real Wage

Capital Intensity

Labor Intensity

Energy Intensity

Table 4. Reference Run: Equilibrium changes in output,prices, and factor intensities induced by depletion.N.B. Factor intensity is defined as factor stockrelative to output.

NationalAverage

GoodsSector

-7%

7%

-3%

8%

-81%

-10%

10%

2%

11%

-80%

CapitalSector

-10%

11%

0%

7%

-80%

EnergySector

-82%

548%

396%

447%

0%

.- -

38

III

D-3216 39

In summary, the differences in the equilibrium impact of depletion

on GNP between the Elephant-Rabbit and ENECON models are due to two major

factors. First, energy and capital production are exogenous to the

Elephant-Rabbit model; they are endogenous to the ENECON model. The stock

of capital in each sector depends on the cost of capital and the target

rate of return demanded by investors. Since capital production depends on

energy in ENECON, the price of capital rises relative to labor as depletion

raises energy prices. As a result, capital is reduced relative to labor,

reducing equilibrium output. In the Elephant-Rabbit, capital is fixed

regardless of the rate of return earned.

Second, the indirect factor requirements of capital and energy

production are captured in the ENECON model. These reinforcing

relationships magnify the direct impacts of depletion but are not captured

in the Elephant-Rabbit model

Sensitivity analysis was performed to verify the analysis above

and gauge the contribution of the indirect energy and capital requirements

to the reduction in GNP. Table 5 presents the equilibrium GNP from the

reference run compared to the value from runs in which the three

reinforcing feedbacks shown in Figure 7 are broken.

D-3216

Equilibrium real GNPSimulation Conditions (percent of initial value)

1. Reference Run 90%

2. Energy not required forenergy production 91.6%

3. Capital not required forenergy production 91.4%

4. Capital not required forcapital production 92.5%

5. Energy and capital not requiredfor energy production; capitalnot required for capital production 94.7%

Table 5. Effect of eliminating the indirectfactor requirements of the energy sector.

The dependence of the energy sector on energy accounts for roughly

16% of the impact, while the energy sector's capital needs (which in turn

depend on energy) account for 14%. The capital sector's capital

requirements contribute 25% of the total. When all three reinforcing

40

III

D-3216 41

relationships are eliminated, the impact of depletion is reduced by nearly

50%.* The remainder of the impact is due to the assumption that capital

intensities adjust to yield the target rate of return for investors.

A second set of sensitivity tests were performed to determine the

effect of the initial capital and labor intensities on the impact of

depletion. Table 6 demonstrates that the equilibrium reduction in real GNP

caused by depletion is not sensitive to the value shares of capital and

labor. Reducing the value share of capital in the goods, capital, and

energy sectors by 66%, 50%, and 33% respectively, adds only two percentage

points to the equilibrium GNP. Increasing the value share of capital has a

correspondingly small effect.

* Due to interactions between the three reinforcing loops, the combinedeffect is less than the sum of the individual effects.

D-3216

Value Share of Factorsof Production

EquilibriumReal GNP(Percent ofinitial value)

1. Reference Run:

Capital Labor

Goods SectorCapital SectorEnergy Sector

.30

.40

. 60

.675

.565

.250

2. Reduced Capital Intensity:

Capital Labor

Goods SectorCapital SectorEnergy Sector

.10

. 20

.40

.875

.765

.450

3. Increased Capital Intensity:

Capital Labor

Goods SectorCapital SectorEnergy Sector

. 40

.50

.70

.575

.465

. 150

90%

Energy

.025

.035

.150

92%

Energy

.025.035.150

89%

Energy

.025

.035

.150

Table 6. Sensitivity of final GNP to capital andlabor intensities.

42

111

D-3216 43

IV.2 The Importance of Substitution

To test the sensitivity of the results shown in Figures 4-7 to the

substitution potential of the economy, the model was simulated under the

assumption that no substitution of capital and labor for energy is

possible. In Figure 8, energy and output march in lockstep: if energy use

declines 1%, output must decline 1%.

. I . . . . . . . . . I . . . . . . . . . I

I I I

I I I

I I I

I I II I. I

I -.?-I -~~~~~~~ I I~~

I-1 I II II I I

I I I

I . I

. . . . . . . . . . . . . . . . . .I I I

I I I

I I I

I I II . I

III~~~~~~~~~~~~~

III~~~~~~~~~~~~~~

III~~~~~~~~~~~~~~

,' _' __

Figure 8. Fixed energy/GNP ratio: effect ofeliminating substitution for energy.

14

. . I . . . . . . . .

II

I

I

I

0 u- ~2

1u I? tt- -;

'·1N'tn ~?

U; - UC'

REAL GP

l

.

I

. . . . . . . . . ,

III

. . .. . . . .

CDNSLM I~

PRICEIEXD

I

I

I

I

I

I

I

o o0

& I · 6c U

>- D I

. - .

. . .. . .. .

I

I

I

I

~~ sl~~~lq~~ · · I~~

llQ YC~~~~~~~~~~~~

~~~lr ~~~~~~~~~91~~~

I

I

I

I

I

-4 zo-

D-3216

Gross national product declines sharply to only 60% of its

original value before recovering, after a century, to 78% of the starting

level. Consumer prices overshoot their final equilibrium, exceeding 160%

of the initial price level for fifteen years. Unemployment also rises

dramatically to almost 25% before returning to normal values after the

120th year. Comparing Figures 4 and 8 reveals that the impact of depletion

is sensitive to the potential for substitution. The difference is most

severe during the adjustment period: in equilibrium, GNP with no

substitution for energy is only 12 percentage points less than in the

reference run and unemployment is the same. In contrast, the depression

created by depletion during the adjustment period reduces GNP by nearly 20

points more than the reference run, boosts unemployment from 15% to nearly

25%, and delays the recovery to equilibrium by nearly two decades.

Because substitution is not possible in the second simulation, the

fraction of national income devoted to energy production rises very sharply

as energy prices rise and a much greater fraction of total investment is

diverted to energy industries. The inflexibility of energy demand forces

the real price of energy up faster and higher than in the base run

(Figure 9). As expenditures for energy rise, the capital sector is less

44

III

45

IO · ·..... ·.. ·I . . . ...... .I . .I .......

I I !

I I I

I I

I I I

I I I I

I I I I ;

' ,,,+ '' REALPRICE:~ ~ ~ ' \ OF ENERGY' ' / ': _ I I I n

I 'r -

I ' /I t

= I I I

I . I I

I XI : a

I t1 AS t

I I

I . Io_ I I O

O

* - 0° YEARS

Figure 9.

.I . .II

I

I

I

I

I

I

I

I

I

I

Zt I

I

I

I

I

Io

III

10

. . . . . . . I

I

I

I

III. . . . . . .

I

I

4}

I

1. . . . . . . II

o

oI.. . . . ..~~~~~~~~~~~~~~~

... . . ..;~~~~~~~~~~~~~~~

I

Fixed energy/GNP ratio: Real price of energy(ratio to initial value).

able to expand, thus making capital still less available.* The capital

shortage constrains the expansion of the energy sector, causing a much

larger dependence on energy imports and subsequent loss of GNP as current

output is used to pay for OPEC oil. As in the base run, the mutual

dependency of the capital and energy sectors, their need to bootstrap

themselves and each other to higher levels of activity, triggers a set of

vicious circles that delay the adjustment to depletion and worsen its

impact.

* The assumption of a fixed money supply may affect the magnitude of thecapital shortage in the short run, but since the problem is ultimately areduction in the real resources available for investment, the long-runresults would not change significantly were the money supply allowed toexpand. In that case, inflation would be added to the other problemsfacing the economy. See Sterman (1980).

D-3216

II

D-3216

A related set of positive feedback loops worsens the problem by

forcing the prices of energy and capital to overshoot their equilibrium

level. Figure 10 shows how depletion can trigger vicious circles causing

increases in the prices of energy and capital: As depletion raises the

price of conventional sources, the energy costs of investments in new

energy technologies rise, raising total energy prices further. Rising

energy prices also raise the production costs and thus the price of capital

plant and equipment, further increasing the costs of energy production.

And as the prices of capital plant and equipment increase, the costs of new

investment in plant and equipment grow as well, adding a third vicious

circle. These positive feedback loops will sustain price increases in

energy and capital and help cause the real price of energy in the second

simulation (Figure 10) to overshoot its equilibrium value by more than 30%.

The price overshoot is more severe in the second run than the base for

without substitution an important set of forces that acted to control the

vicious circles is lost.

The strength of the feedback loops shown in Figure 10 depends on

the energy and capital intensities of energy and capital industries. The

more energy- and capital-intensive these industries are, the stronger the

vicious circles. The dramatic rise in the real costs of technologies such

as coal liquefaction, shale oil, and nuclear power can be traced, in part,

to their massive energy and capital requirements. Figure 10 suggests

energy policies based on these technologies such as a government-sponsored

synfuels program may be self-defeating. By triggering further price

increases in capital and energy, a synfuels program may be plagued by cost

46

D-3216

ENERGY RESOURCES -

ENERGY PRODUCTION

IENERGY tPRICES "'

±it

+ PRODUCTIVITY OFENERGY SECTOR

ICOST OF ENERGYPROn KIT ON

COST- OF CAPITAL PRICE OF CAPITALPLANT AND --- PLANT ANDEQUIPMENT +EQUIPMENT +

+ -

CAPITAL REQUIREMENTSOF ENERGY SECTOR

DEMAND FOR CAPITALPLANT AND EQUIPMENT

Figure 10. Depletion triggers vicious circles causingprice increases in energy

11

47

f -F�

-------- --- -

..------ II'i +

D-3216

overruns, delays, and reduced profitability. Added government subsidies

would be required to complete the program. In contrast, an energy policy

based on conservation and renewable technologies with relatively low

capital and energy requirements would be less likely to trigger the vicious

circles. By reducing the overall dependence of energy and capital

production on capital and conventional energy, the strength of the vicious

circles will be reduced as well, thus helping to protect the economy from

unexpected increases in energy costs.

IV.3 The Importance of Delays

The second notable aspect of the reference run was the long

adjustment period required before equilibrium is re-established. During

the adjustment period, GNP is reduced below its long-run equilibrium,

consumer prices rise above their final value, and unemployment more than

triples. Indeed, the transient consequences of depletion are much more

important than the equilibrium consequences.

Long delays cause the slow adjustment. There are very long delays

in replacing the original capital plant with energy-efficient capital and

adjusting production technologies to higher energy prices. In addition, as

depletion proceeds, the demand for capital grows, but it takes time for the

capital sector to expand sufficiently to satisfy the extra demand as well

as supply the extra capital it needs for itself. Workforce participation

adjusts slowly to changing wage and employment conditions, a reflection of

the difficult lifestyle changes implicit in the decision to work for

48

III

D-3216 49

yourself (as a housewife or subsistence farmer, for example) or enter the

workforce (as a factory or farm worker).

The delay in substituting new capital for old is one of the most

important sources of the more severe intermediate-term impact. In the base

run, two decades (the average life of capital) are required to replace old

capital with energy-efficient capital. Thus, even though the price of

energy reaches its equilibrium value by about the 60th year of the simula-

tion, the energy/GNP ratio continues to decline for half a century before

reaching equilibrium (Figure 6). Because of the long life of capital,

energy demand is relatively inelastic in the short run, resulting in a

transient increase in the fraction of national income devoted to energy

production. Figure 11 shows that expenditures on energy as a fraction of

GNP rise in the reference run from nearly 6% to a peak of more than 15%

before settling back to the original level. The fraction of total

investment going to the energy sector more than doubles in the same period.

Expenditures are sustained above normal for a period of 80 years as old

capital is slowly replaced with energy-efficient plant, equipment, and

housing. Thus, the delay in replacing existing capital directly

contributes to the length and severity of the adjustment period by reducing

the discretionary income of households and firms.

· I

I

III

II

I

I

I

I

I

I

.I

I

. .. . . .. .. I . .. . . . ..I

I

I

I

I

I

I

I

I

I

I

I

EPENDITUES ONENERGY AS AFRACTION OF GP

I I

I I

I I

I I

I B

I I

I I

I I

I I

I I

I I

I I

I I

I I

I BI I~~~~~~~~~

I I B B I B

I I B B I I0' I * ' ·· · · ·I - · · " ·- * . ·· · ·I . · ·. . . . . .

0 0 0 0 0__r OD YEARS I 0oao~~~~~~~~~~~~~~~~~~~~Cg o g o g~~~~~~~~~~~~~~~~~~~~~n

Figure 11. Reference Run: Productive Resources of the economyare diverted into the energy sector due to delays insubstitution.

The reduction in discretionary income worsens the intermediate-

term impact in yet another way. As productive resources are diverted to

the energy sector, the ability of the economy to purchase needed capital is

depressed. Further, the ability of the capital sector to purchase the

capital it needs is reduced. Reduced investment and availability of

capital delays substitution of energy-efficient capital for old, sustaining

energy consumption and worsening depletion. A vicious circle is started

(Figure 12) in which high energy costs reduce the ability of firms and

households to substitute away from energy, worsening depletion and pushing

D-3216 50

'I IIIIIr

III

I

LAU11

I

I

I

I

I

I

I

I

I

I

I

III

D-3216 51

energy costs up still faster. By reducing the resources available to

invest in efficient capital, depletion undermines the very forces acting to

control it, causing a further decline in well-being. The vicious circle

shown in Figure 12 stretches out the adjustment period and causes the price

level, real price of energy, and unemployment to overshoot (and GNP to

undershoot) their long-run values.

-- INVESTMENT IN ENERGY E--NERGY DEMANDRGY DEMAN

+

+I

TOTAL ·INVESTMENT-

"+--- --- REAL INCOME

Figure 12. Rising energy prices reduce real income and investmentin energy efficient capital, delay substitution, andmaintain energy demand, causing further price increases.

r-r1-Tr1Trr ADfrnlr-1-I- I , C1I I %, F I ML

rf-"wL "' T/"Ir"

k-t-J MtKUT PRICES

D-3216 52

Reducing the magnitude of the intermediate-term downturn and

smoothing the transition becomes an extremely important area for energy

policy. To determine the role of the delay in reducing the energy

intensity of production, the model was simulated under the assumption that

energy intensity can be reduced twice as fast as in the base run, or in

half the lifetime of capital. Such an assunption might reflect an

intensive conservation and retrofit program. The long-run elasticity of

substitution is unity, as in the base run. Figure 13 compares GNP and

unemployment in the two runs.

0 0

II

. . . .. . . . . . . . . . . ·. . . . . . . .· . . . . . . . . . . . . . . . ·. . .e~- i s

BASE RUN

I I I I ACCELERATED RETROFITAM COINSERVATION

I I I I*c'J0 · · · · ·· · .· · .·· .· ..· .'. .· ·· ·· · · ·· .

0.I.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I , I I I

I 1 ~~ ~I I I

O~(U.c~ J

I I I I

II I I I I

00..... . . I..... ..

0o 0 EA0 0t ao Iv y, EARS N '0 0

Figure 13. Effect of reducing delays in improving energyefficiency on GNP (percent of initial value)and unemployment (percent of workforce).

D-3216 53

The final equilibrium is the same, but faster adjustment of energy

intensity to energy prices significantly reduces the undershoot of GNP and

overshoot of unemployment. More rapid conservation lowers the fraction of

resources devoted to the energy sector in the short term, thereby reducing

the strength of the vicious circles that drive output below its long run

value.

D-3216

V. AN EXAMPLE OF ENERGY POLICY ANALYSIS

Energy policy has been hampered by the inability of existing

models to assess the macroeconomic impacts of policy initiatives. The

macroeconomic effects of policies such as price controls, rationing,

subsidies for synthetic fuels, taxes, and import quotas are poorly

understood. Many policies generate forces that oppose or undermine the

intent of the policy. For example, government subsidies for production of

synthetic fuels may increase the supply of oil, but in order to finance the

program the government must either raise taxes, cut spending in other

areas, or increase the deficit. The first two choices benefit the energy

sector at the expense of the rest of the economy; the last contributes to

inflation, thereby reducing the real value of the subsidy and causing a

need for even higher subsidies.

A tax on energy use with a fully compensating income tax reduction

was chosen to demonstrate the ability of the ENECON model to evaluate

energy policies in the context of the economy as a whole.

A tax on energy use directly affects incentives for energy con-

sumption and production; rebating the proceeds by reducing tax rates

creates incentives in other sectors of the economy to rebalance the mix of

productive factors and the composition of final demand. Thus, such a tax

provides a good test of the model's ability to capture important macro-

economic and social feedbacks.

· YYL __m _YI ___~I *__I _·i~ -- ^- --(-------

54

III

D-3216 55

The tax proposal tested in the model consists of a fixed BTU

surcharge on energy use matched by a reduction in both corporate and

personal income tax rates sufficient to offset the revenues generated by

the tax. Thus, government revenues are initially unchanged. The reduced

income tax rates allow consumers to purchase the same market basket of

goods as before the tax, though, of course, it is to their advantage to use

less energy-intensive items. Similarly, firms will be able to achieve

roughly the same after-tax return on investment despite higher unit costs.

If large enough, the tax would create strong incentives to reduce energy

use and increase use of other factors, particularly labor.

Several simplifying assumptions have been made in representing the

tax. First, it is imposed in one step, while a phased introduction would

probably be required in reality. Second, income tax rates are held

constant at the new, lower level; as energy use (hence tax revenues)

declines, government spending is assumed to drop correspondingly. Third,

the tax is levied on all energy; in reality only oil and gas would be

taxed, thus stimulating production of coal and renewable alternatives.

The initial conditions are the same as those used to produce the

reference run. In the 40th year of the simulation, a BTU tax sufficient to

cut both personal and corporate income tax rates by 20% is imposed. The

result, shown in Figure 14, is a transient reduction in GNP and increase in

unemployment relative to the base case followed by a long-term adjustment

D-3216 56

>- 5

.,.v-

I ISE II BASERL I

I I I

.o,

U(:3..N%

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .0

c0 YEARS gN g 0

Figure 14. Effect on GNP and employment of an energy tax withcompensating reduction in income taxes (GNP as apercent of initial value; unemployment as a percent ofthe workforce.)

to an equilibrium GNP some 10% higher than the base. In equilibrium, real

GNP is restored to its original level. The long-run effect of the tax is

to increase real GNP (and consumption) by reducing energy demand,

increasing capital intensity, and increasing workforce participation. The

I

I

I

I

I

I

III

D-3216 57

reduction in energy demand frees labor and capital from the least produc-

tive sector for use in other sectors. Capital intensity increases

becausethe after-tax cost of capital services is reduced. Workforce

participation increases because the after-tax wage has increased;

employment increases because labor becomes relatively less expensive than

energy.

The cause of the transient reduction in GNP lies in the pricing

policies of the production sectors. In the model, prices are determined by

both unit costs and a traditional profit margin or markup on unit costs,

along with supply and demand pressures. Even though reducing the corporate

tax rate allows the same after-tax return on investment despite higher unit

costs, it will take time for managers to become accustomed to the lower

markup required to achieve the target after-tax return. Before the

adjustment to a new markup level is made, prices are pushed above their

long-run equilibrium level, reducing real income, aggregate demand, and

employment.

The effect of traditional rules of thumb for markup on prices and

GNP illustrates the importance of capturing managerial decision making in a

model. By including a representation of the price setting process that

does not rely on equilibrium assumptions, a potential source of managerial

resistance to the policy initiative has been identified. The tax policy

can then be improved by designing ways to overcome such institutional

resistance.

D-3216 58

The effect of the tax policy on GNP under the assumption of a

speedy adjustment to the new markup levels is shown in Figure 15. Here

the long-run benefit is the same, but the transient decline in GNP below

the reference run is eliminated and the transient increase in unemployment

is substantially reduced.

0c.or

.t, I I I I

' & BASE RUNI I I I

: J[~riGY TAX WITH REDUCTION IN

I r ~~~~~I : INCOME TAXES AM)N*I~~~ ~~~QUICK MARKUP

.· .. ..u.'r. ' .. .o · ·· · · ··

oI · ···

··

· ··· ·

·· ° DLUSWENT °

Lu

!\

' ~~~I I I I I-

C1 . . . . . . . . . . . . . . . . .~~~~. . . . . . . I I

oo I · · ·· I ··· I · · * * * · ,, ·· I00

0 gco YEARS

Figure 15. Effect of energy tax with compensating income taxreduction and swift adjustment of profit margins.

III

D-3216 59

VI. CONCLUSION

The simulations presented above reveal several important results.

Energy depletion does have a significant effect on GNP, consumption, and

employment. Even though the energy sector is a small part of the economy,

depletion can significantly reduce economic growth. The impact of

depletion is particularly severe in the intermediate term, before the

economy can fully adjust to higher energy prices. During the adjustment

period, depletion can dramatically reduce national output and create

massive unemployment.

By diverting resources to the energy sector, depletion reduces

non-energy output. The indirect requirements of the energy sector,

particularly the higher level of capital required, worsen the impact since

labor and capital must be diverted to produce more capital for use in the

energy sector. The capital and energy requirements of the capital and

energy sectors further reinforce the indirect effects.

The greater the potential for substitution, the milder the impact

of depletion on GNP. If substitution is difficult, the demand for energy

will be very inelastic; a much larger fraction of GNP must then be devoted

to energy production and capital production for the energy sector. But

while the potential for substituting labor and capital for energy

influences the long-run level of output somewhat, the intermediate-term

impact of depletion is particularly sensitive to the substitution

potential. The surge in demand for capital during the adjustment period

D-3216

can create a capital shortage and constrain the ability of the capital

sector to increase its own capacity, aggravating the shortage further.

Long delays in capital turnover cause expenditures on energy and

investment in energy to rise more in the short run than in the long run.

The diversion of extra resources to the energy sector reduces GNP and the

standard of living below their long-run values during the period of

adjustment. Thus the disequilibrium effects of depletion are quite

significant.

Speeding conservation efforts can substantially smooth the tran-

sition and reduce the undershoot of GNP by reducing the fraction of GNP

devoted to energy and weakening the vicious circles that reduce the ability

of firms and consumers to cope with higher prices.

The results presented here illustrate the importance for energy

policy of the interactions between energy depletion and macroeconomic

behavior. Though the results are preliminary, several important areas of

concern have been identified, and the feasibility of an integrating

framework for energy-economy analysis has been demonstrated. Future model

development will address the simplifying assumptions made in the

simulations; future policy analysis will integrate the results above with

results on inflation, monetary and fiscal policy, and other energy policy

initiatives.

60

D-3216

APPENDIX

EQUILIBRIUM STRUCTURE OF THE ENECON MODEL

The ENECON model is a dynamic model. The model is designed

to reflect actual managerial and personal decision-making by firms and

households. In reality, information about the future is not available

for current decision making. Indeed, information about the current

state of affairs is often unavailable. Important perception delays

exist; expectations are often wrong; goals are often unrealized;

conflicting pressures may arise, forcing compromise.

Thus, in the model, information about the future is not used

in current decisions. Perception delays are included where

significant, for example in estimating the "true" demand for a

product, factor availability, consumer incomes, and the relative

productivity of different factors of production. Expectations are

formed in the model on the basis of past behavior and trends; there is

a delay in forming expectations since it takes time to separate an

underlying trend from noise and short-term fluctuations. Goals, such

as the desired capital stock or desired energy intensity of

production, cannot be realized instantly: capital must be ordered,

built, and delivered; old, energy-intensive capital must depreciate

before it can be replaced with energy-efficient capital (except for

retrofits). Conflicting pressures may delay or prevent a goal from

61

D-3216

being reached. For example, higher energy prices may increase desired

capital or desired labor; at the same time, higher costs may create a

cash crunch preventing firms and consumers from ordering as much

capital or hiring as many people as they need.

Because the model captures the important pressures and delays

that characterize the real economy, there is no guarantee that the

simulated system will be in equilibrium at any given time. Indeed,

the simulations above indicate the disequilibrium behavior of the

economy is extremely important for energy analysis and policy.

However, if undisturbed, the system will eventually reach a stable

equilibrium.

Equilibrium in the context of the simulations presented above

means desired states and actual states are equal. Further, since

secular growth has been excluded, all state variables will be constant

in equilibrium. For example, investment will just balance

depreciation, income will equal expenditures, net saving will be zero,

and net borrowing will equal net debt retirement plus defaults;

desired factor stocks and actual factor stocks will be equal, money

balances will equal their desired levels, and return on investment

will equal the required rate of return.

The description below presents an overview of the conditions

that determine equilibrium in the model and is not intended to be

complete documentation of either the equilibrium conditions or the

62

III

D-3216 63

dynamic structure of the model. Full documentation will be available

at the completion of the work described here.

The equilibrium reached by the model conforms to neoclassical

theory. Producers are assumed to be profit-maximizers; consumers are

assumed to be utility-maximizers. The equilibrium conditions fall

into two broad categories: those that determine the real equilibrium

and those that determine the price level. The work reported here

focuses on the real effects of depletion; for a description of the

determinants of the price level, see Sterman (1980).

The determinants of the real equilibrium fall into three

classes: first, conditions that determine the optimal mix of factors

of production within each sector; second, conditions that determine

the return on investment and price of output in each sector; and

third, conditions that balance the demand for factors with the supply

throughout the economy. The three sets of conditions are described

more fully below.

D-3216

A.1 Factor Balance Within Each Sector

The heart of each production sector is a neoclassical

production function (the household employs an analogous utility

function). In equilibrium the marginal revenue product of each factor

equals its marginal cost:

(1) MRs*MPf,s/MCf,s 1 all f,s

(2) MRs = Ps

(3) MCfs Pf for f = labor, energy

(4) MCfs Pf CCRs for f = capital

where,

MRs = Marginal Revenue in sector ($/output unit)

MPf s = Marginal Productivity of factor in sector

(output units/year/factor unit)

MC = Marginal Cost of factor in sectorf,s

($/year/factor unit)

P = Price of output in sector ($/output unit)sPf = Price of factor ($/factor unit)

CCR = Capital Charge Rate (%/year)

f = index of factors of production

s = index of production sectors

Both output and factor markets are assumed to be competitive,

thus marginal revenue equals the current price of output and the

marginal cost of a factor equals its current price. The assumption of

competitive output and factor markets does not imply these markets are

--- --'-�--~-"I-�^�-�I�"----�--s�^l--

64

III

D-3216 65

always in equilibrium or that, for example, the wage always equals the

marginal revenue product of labor. Instead, competition here reflects

the assumption that each firm in a sector perceives itself to be small

relative to the sector as a whole and unable to influence factor or

output prices.

In the simulations presented above, a standard Cobb-Douglas

production function with constant returns to scale was used.

Cobb-Douglas functions with constant returns to scale have an

elasticity of substitution of unity between factors.*

In the energy sector the production function is modified by

a term representing the impact of depletion. The isoquants of the

energy sector's production function shift downward as resources are

depleted:

(5) Qe = D*Q(K eLe'Ee)

(6) D = D(Resources)

* In the simulation with no substitution of capital and labor forenergy (Section IV.2; Figures 10-12) the energy intensity ofproduction in each sector was fixed at its initial level throughoutthe simulation preventing the rising price of energy from reducingenergy intensity. Thus, the energy/output ratio in each sectorremained constant, implying no substitution of capital or labor forenergy was possible.

D-3216

where

Qe = energy production (BTU/year)Q(K,L,E) = energy sector production function

D(resources) = depletion function (dimensionless)

Depletion is assumed to be Hicks-neutral; the depletion function is

monotonic and reaches an asymptote at the backstop level.

A Cobb-Douglas utility function is used in the household

sector. The optimal factor mix is determined by the following

conditions:

(7) MVUh*MUf,h/MCfh = 1

TF(8) MVUh [ F *MC ]/U

fh h f,h hf=1where

MVUh = Marginal Value of Utility ($/utility unit)

MUf h = Marginal Utility of factor in household

(utility units/year/factor unit)

MCf h = Marginal Cost of factor in household ($/factor unit)

U = Utility (utility units/year)

F = Factor of utility (factor units)

h = household sector

f = index of factors of utility

TF = Total Factors of utility

The household adjusts its factor mix until the marginal

dollar value of the utility contributed by the factor equals its

66

III

D-3216 67

marginal cost. The marginal dollar value of a unit of utility is the

total cost of the current level of utility at the margin divided by

the current level of utility. These conditions ensure that in

equilibrium the household employs the optimal mix of factors of

utility. Given the optimal mix, the equilibrium stock of each factor

is determined by the income of the household: in equilibrium, the

cost of maintaining the stocks of factors will just exhaust household

disposable income, insuring that utility is maximized.*

A.2 Return on Investment and Price

It is assumed throughout that investors require a certain

after-tax return on investment. The target return includes the real

interest rate plus a risk premium to compensate investors for

undertaking the investment. In equilibrium, the net income of the

production sectors must yield the target return to all investors.

Ignoring taxes, investment tax credits, and debt for simplicity:**

(9) GROI = NIs/MVCs

(10) NIs PsQs(Ws*Ls+Pe*ES+Pk*Ks/ALCS)

(11) MVCs = PkKs

* Since the model reaches a stationary equilibrium involving nogrowth, equilibrium net savings equals net investment equals zero.Therefore, the entire disposable income of the household is spent onconsumption.

** Though taxes, investment tax credits, and debt are included in themodel.

D-3216 68

where

GROI = Goal for Return on Investment (%/year)

NI = Net Income ($/year)

MVC = Market Value of Capital ($)

P = Price of output ($/unit)

Q = Quantity of output sold (units/year)

L = Labor force (persons)

W = Wage ($/person-year)

E = Energy consumption (BTU/year)

Pe = Price of Energy ($/BTU)

-Pk = Price of capital ($/capital unit)

K = Capital Stock (capital units)

ALC = Average Life of Capital (years)

The equilibrium condition for price is simply a restatement

of the condition for return on investment. In equilibrium, prices

just cover the costs of production and normal profit (determined by

the target return on investment and including depreciation). Again,

ignoring debt, taxes, and investment tax credits:

(12) P = UC + PMs s s

(13) UCs = (Ws*Ls + Pe Es)/QS

(14) PMs = [(GROIS *MVCs) + DEPs]/Qs

(15) DEPs = PkKs/ALCs

where

P = Price ($/unit)

UC = Unit Costs ($/unit)

PM = Profit Margin ($/unit)

DEP= Depreciation ($/year)

III

D-3216 69

In the short run, imbalances between supply and demand can

raise or lower price above the equilibrium level. But supply/demand

imbalances cannot exist in equilibrium (unless price controls or other

market distortions are introduced). For example, higher than

equilibrium prices will depress demand and induce higher profits and

hence entry of firms into the sector, increasing supply and forcing

price down to the equilibrium level. Lower than equilibrium prices

will result in exit of firms, reduced supply, and a restoration of

equilibrium.

Wage determination differs from other factor prices. Each

sector generates its own wage, and competes for labor against the

other sectors. Five factors influence the level of wages in each

sector: return on investment in each production sector relative to

the target return, liquidity in each sector, the relative availability

of labor in each sector, wages in other sectors, and the perceived

rate of inflation in consumer prices. In equilibrium, ROI will equal

its target value, liquidity will be normal, labor availability will be

normal, wages in all sectors will be equal, and consumer prices will

be stable. Imbalances in the relative availability of labor or in the

relative wage will cause wages to adjust, employment to change, labor

to migrate between sectors, and workforce participation to vary until

the imbalance is eliminated. The equilibrium wage will equal the

marginal revenue product of labor.

D-3216

A.3 Supply and Demand Balance throughout the Economy

The equilibrium conditions described above ensure that each

sector chooses the optimal mix of factors and earns an adequate

return. The conditions described below determine the overall level of

output and the general equilibrium for the economy.

First, the total consumption of each type of output must

equal production:

(16) Energy: Qe Eg+Ek+Ee+Eh

(17) Capital: Qk K /ALCg + Kk/ALCk + Ke/ALCe + Kh/ALCh

(18) Goods: Qg Gh/ALGh

where

Q = production (output units/year)

E = energy consumption (BTU/year)

K = capital stock (capital units)

ALC = average life of capital (years)

G = goods stock (goods units)

ALG = average life of goods (years)

e,g,k,h = energy, goods, capital, household sectors

70

III

D-321.6 71

Second, household expenditures must equal income. Ignoring

taxes and debt, the household spends money on goods, capital, and

energy; it receives wages and profits from the production sector as

income:

(19) PgGh/ALGh + PkKh/ALCh + PeEh = W(Lg+Lk+Le) + PkGROI(Kg+Kk+Ke)

Finally, the sum of the labor employed in the production

sectors and the labor "employed" in the household must equal the total

working age population Lt:

(20) Lt = Lg+Lk+Le+Lh

Together, the three sets of equilibrium conditions determine

the real equilibrium of the simulated economy. The equilibrium

reached is a general equilibrium since all markets are simultaneously

clearing and there are no exogenous factor or output prices. Only the

total working-age population is given; the allocation of labor among

the different sectors and the various production rates are endogenous.

A.4 Initial Conditions and Parameters

The model has been initialized to correspond to economic

conditions in the post-war era and before the energy price increases

of the 1970s. The year 1950 was chosen as a reference year for GNP

and energy production. Since the purpose of the model is policy

analysis and not prediction, econometric estimation has not been

performed at the current stage of development. Parameters and initial

D-3216

conditions were chosen to be consistent with aggregate data and

economic theory; sensitivity analysis was performed to test the

importance of errors in parameter choice (see Sections IV.1, IV.2, and

IV. 3).

Table 7 presents some of the more important parameters and

initial conditions. Where applicable, actual values for 1950 are also

given.

Variables

GNP2

Energy Production

Population

Unemployment

Real Interest Rate

Target after-tax ROI

Capital Charge Rate

Raw Corporate Tax Rate

Average Personal

Income Tax

Average Life of Capital

Average Life of Goods

Model Value1

$254 billion 3

33.4 quads/year

151 million

5%

4%/year

8.5%/year

19%/year

50%

1950 Value

$285 billion4

33.6 quads/year5

151 million 6

5.1% 7

30%

20 years

2 years

NOTES

1. Model values refer to the reference run.2. 1950 dollars3. The model is initialized for a stationary equilibrium. Since the

real economy was growing, model-generated investment (and GNP) areless than the actual values.

4. Dept. of Commerce (1975) p. 2245. DOE/EIA (1978) p. 76. Dept. of Commerce (1975) p. 87. Ibid., p. 127

Table 7. Selected Initial Conditions and Parameter Values

I� ___�_I_�_��_�__

72

III

D-3216 73

The value shares of each factor of production (the exponents

in the Cobb-Douglas production functions) are presented in Table 8.

The exponents for capital and labor were chosen to reflect the

assumption that the energy sector is more capital intensive than the

capital sector which, in turn, is more capital intensive than the

goods sector.

Capital Labor Energy Goods

Goods Sector .30 .675 .025 0

Capital Sector .40 .565 .035 0

Energy Sector .60 .250 .150 0

HouseholdSector .13 .200 .020 .65

Table 8. Value shares of factors of production

The exponents for energy were chosen to be consistent with

energy production in the reference year. The exponents are quite

small in the household, goods, and capital sectors, reflecting the

small contribution of energy to production in the 1950s. The large

exponent for energy in the energy sector requires some explanation.

With a Cobb-Douglas production function, the equilibrium condition for

energy use in the energy sector (equation 1) becomes:

D-3216

(1') PeQeee/PeEe = 1

where Pe = Price of energy

Qe = Energy production

Ee = Energy consumption in the energy sector

ee = Exponent for energy in the energy sector

Solving for ee yields:

(11") ee = Ee /Qe

In 1950, 5.0 quads/year out of the gross energy production of

33.6 quads/year were consumed in the generation of electricity

(DOE/EIA 1978, p. 7). Thus, the value for ee becomes:

ee = Ee/Qe = 5.0/33.6 = .15

This is probably low since it ignores energy consumed in other forms

of energy production such as oil exploration, refining, and

distribution; coal mining; and electricity transmission and

distribution losses.

74

III

D-3216 75

REFERENCES

Backus (1977) George A. Backus, "FOSSIL 1:Documentation". Prepared for ERDAby the Dartmouth System DynamicsGroup, Hanover, NH: Thayer Schoolof Engineering (1977).

Baughman and Joskow (1975)

Berndt and Wood (1977)

Berndt and Wood (1979)

Carasso et al. (1975)

Cazalet (1977)

Martin L. Baughman and Paul L.Joskow, "Energy Consumption and FuelChoice by Residential and CommercialConsumers in the United States."MIT-EL-75-024, Cambridge, MA: MITEnergy Laboratory (1975).

Ernst R. Berndt and David O. Wood,Consistent Projections of EnergyDemand and Aggregate EconomicGrowth: A Review of Issues andEmpirical Studies, MIT Energy Labworking paper MIT-EL 77-024WP,Cambridge, MA (June 1977).

Ernst R. Berndt and David O. Wood,"Engineering and EconometricInterpretations of Energy-CapitalComplementarity", American EconomicReview, 69, 3, June 1979.

Meir Carasso et al., "The EnergySupply Planning Model", Vol. 1 & 2,PB-245, 382, San Francisco, CA:Bechtel Corp. (August 1975).

Edward G. Cazalet "GeneralizedEquilibrium Modeling: TheMethodology of the SRI-Gulf EnergyModel", Palo Alto, CA: DecisionFocus, Inc., (May 1977).

D-3216

Department of Commerce (1975)

DOE (1979 )

DOE/EIA (1978)

DOE/EIA (1979)

Department of Commerce, Bureau ofthe Census, Historical Statistics ofthe United States, Colonial Times to1970, Part 1. Washington, DC: Bureauof the Census, (1975).

Department of Energy, NationalEnergy Plan II, DOE/TIC 10203 andAppendix A: World Oil Prices,DOE/TIC-10203 (APPA), Washington DC:Department of Energy (May 1979).

Energy Inforamtion Administration,Annual Report to Congress: Vol. III,1977, Statistics and Trends ofEnergy Supply, Demand, and Prices,Washington, DC: Dept. of Energy,SP-AN74/A(77), (May 1978).

Energy Information Administration,U.S. Dept. of Energy, Survey of theResearch into Energy-EconomyInteractions, Vol. 1- Survey; Vol.2-Annotated Bibliography,Washington, D.C.: US GovernmentPrinting Office, (April 1979).

Forrester (1961)

Forrester, Massand Ryan (1976)

Forrester and Mass (1979)

Jay W. Forrester, IndustrialDynamics, Cambridge, MA: MIT Press(1961).

Jay W. Forrester, Nathaniel J. Mass,and Charles J. Ryan, "The SystemDynamics National. Model:Understanding Soci o-EconomicBehavior and Policy Alternatives,"Technological Forecasting and SocialChange 9, 51-68 (1976).

Jay W. Forrester and Nathaniel J.Mass, "Response to the Library ofCongress Analysis of the SystemDynamics Energy Proposal", MITSystem Dynamics Group Working PaperD-3165, Cambridge, MA: SystemDynanics Group (November 1979).

i-----^-� -- I

76

D-3216

Charles J. Hitch, Modeling Energy-Economy Interactions: FiveApproaches, Washington, DC:Resources for the Future (1977).

William W. Hogan, "ProjectIndependence Evaluation System:Structure and Algorithms",Proceedings of Symposia in AppliedMathematics, Vol. 21, Peter Lax(ed.), Providence, RI: AmericanMathematical Society (1977).

Hogan (1978) William W. Hogan, "Energy and theEconomy", EPRI EA-620, Vol. , PaloAlto, CA: Electric Power ResearchInstitute (January 1978).

Hogan and Fromholzer (1977)

Hogan and Manne (1977)

Hogan and Weyant (1978)

William W. Hogan and DennisFromholzer, "On Estimating theElasticity of Substitution forPrimary Energy: Documenting NegativeResults with Traditional Methods",Energy Modeling Forum, EMF 4 .1,Stanford, CA: Stanford University(July 1977).

William W. Hogan and Alan S. Manne,"Energy-Economy Interactions: TheFable of the Elephant and theRabbit," Modeling Energy-EconomyInteractions: Five Approaches,Charles J. Hitch (ed.), Washington,DC: Resources for the Future (1977).

William W. Hogan and John P. Weyant,"An Experiment Designed to Determinethe Elasticity of SubstitutionBetween Energy and Other Inputs byMeasuring the Price ElasticityImplicit in Energy Models ," EnergyModeling Forum, EMF 4.O, Stanford,CA, Stanford Univeristy, (March1978).

11

, . I-I I .1 · ·. I -"I I c·r l ..·r' .I \11 l 4 . - ,rrr,

77

Hitch (1977)

Hogan (1 977)

D-3216

M. King Hubbert, "Energy Resources",Resources and Man, Commnittee onResources and Man, San Francisco:W.H. Freeman (1969).

Hudson and Jorgenson (1978)

Hull (1979)

Edward A. Hudson and Dale W.Jorgenson, "The Long TermInterindustry Transaction Model: ASimulation Model for Energy andEconomic Analysis", General ServicesAdministration, GSA/FPA/MCL TR-109,Washington, DC: Federal PreparednessAgency (September 1978).

Everson W. Hull, "MacroeconomicEffects of Alternative Excise Taxeson Domestic and Foreign Oil with aReduction in Federal GovernmentPersonal Tax Receipts Equivalent.tothe Additional Revenues",Washington, DC: Library of CongressCongressional Research Service(September 1979).

Kennedy and Houthakker (1978)

Manne (1977)

Manne (1978)

Mass (1975)

Hendrik S. Houthakker and MichaelKennedy, "Long Range EnergyProspects", Harvard Institute ofEconomic Research Discussion Paper#634, Cambridge, MA:. HarvardUniversity (July 1978).

Alan S. Manne "ETA-MACRO: A Model ofEnergy-Economy Interaction",Modeling Energy-EconomyInteractions: Five Approaches,Charles J. Hitch (ed.), Washington,DC: Resources for the Future (1977).

Alan S. Manne, "The Fable of theElephant and the Rabbit", Paperpresented at AAAS Meeting,Washington, DC: (February 14, 1978).

Nathaniel J. Mass, Economic Cycles:An Analysis of Underlying Causes,Cambridge, MA: MIT Press (1975).

� I__ � �1��1_1_�

78

Hubbert (1969)

III

D-3216 79

Nordhaus (1973)

Parikh (1976)

Reister and Edmonds (1977)

Solow (1974)

Sterman (1980)

William D. Nordhaus, "The Allocationof Energy Resources", BrookingsPaper on Economic Activity, Vol. 3,(1973), 529-570.

Shailendra C. Parikh, "AnalyzingU.S. Energy Options Using the PILOTEnergy Model," Systems OptimizationLaboratory, TR SOL 76-27, StanfordCA: Stanford University Departmentof Operations Research (October1976).

David B. Reister and James A.Edmonds, "A General EquilibriumTwo-Sector Energy Demand Model",Modeling Energy-EconomyInteractions: Five Approaches,Charles J. Hitch (ed.), Washington,DC: Resources for the Future (1977).

Robert M. Solow, "The Economics ofResources or the Resources ofEconomics", American EconomicReview, 64 (May 1974), 1-14.

John D. Sterman, Energy andInflation. Paper prepared forOffice of Analytical Services, U.S.Department of Energy. ReportD-3203, System Dynamics Group,Massachusetts Institute ofTechnology, Cambridge, MA, 02139,(March, 1980).

Sussman and Rousseau (1978)

WEFA (1978)

S.S. Sussman and W. F. Rousseau, "ADemonstration of the Capabilities ofthe Livermore Energy Policy Model,"UCRL-52508, Livermore, CA: LawrenceLivermore Laboratory, University ofCalifornia (July 1, 1978).

Wharton Economanetric ForecastingAssociates, The Wharton AnnualEnergy Model: Development andSimulation Results, Philadelphia:Wharton EFA, Inc. (March 1978).

i

D-3216 80

White (1971) David C. White, "Energy, theEconomy, and the Environment",Technology Review, October/November,1971, 18-31.

_a� � �·_____��___